DISTRIBUTION OF THE CHAETOGNATHA 209 



and in view of the absence of Discovery records to confirm them it seems probable that most of these 

 anomalous records can be regarded as misidentifications of badly preserved material. For example, 

 S. hexaptera has been reported from the Antarctic by Fowler (1907), Ritter-Zahony (191 1), Germain 

 (1913), and Johnston & Taylor (1921). In all these cases, except that of Ritter-Zahony, confusion with 

 specimens of S. gasellae on which the fins were damaged seems to be the most likely explanation; 

 after 1921 the same confusion does not seem to have arisen, although, surprisingly enough, very 

 recently three large specimens of S. hexaptera were reported from the Mediterranean as specimens of 

 S.gazellae (Furnestin, 1953, 1955). In the case of Ritter-Zahony (191 1) there can be little doubt that 

 the specimens reputedly from 66° 2' S 89° 38' E were 5. hexaptera, since he was very familiar with 

 that species and with S. gazellae, but it is significant that in the same haul he found K. suhtilis, another 

 subtropical species, and it is possible that there may have been confusion with a haul made on the 

 same date (22nd November) in 1901 (not 1902 as reported) when the ship's noon position was 34° 9' S 

 17° 26' E in the vicinity of Cape Town where one would expect to find these two subtropical species. 

 S. serratodentata has been reported by Burfield (1930) from the Ross Sea. Only one specimen was 

 taken, and the only explanation I can suggest is contamination from a more northerly sample. It is 

 quite easy for a specimen to be caught inside the lid of a sample jar and the lid to be put on to another 

 jar, or for a single specimen to be left under the rim of a petri dish while sorting. Several authors have 

 recorded the species from the northern part of the Subantarctic where its presence can be confirmed 

 by Discovery records. 



IV. Doubtful species 

 S. neglecta. This tropical species was reported by Johnston & Taylor (1921) from 64° 34' S 1 17° i ' E, 

 and must be regarded as a case of mistaken identification. The species is rare even in subtropical 

 waters and certainly does not occur in Discovery collections from the Southern Ocean. 



VERTICAL DISTRIBUTION 



The vertical distribution of the relative density of the populations of chaetognaths is shown in Figs. 3, 

 5, 7, 8, 10, II, 12 and 13 on three lines of stations, one in the Atlantic sector (meridian of 0°), one in the 

 Indian Ocean sector (meridian of 90° E) and one in the Pacific Ocean sector (meridian of 80° W). 



Before comparing the results from these sections it is necessary to indicate that there are certain 

 obvious physical differences involved. The line in 90° E is perhaps the most simple; it is relatively long 

 and the subantarctic zone extends from about 40° S to about 52° S, both convergences being rather 

 indistinct. The antarctic zone was accessible from about 52° S to the ice edge in 65° S. Only Station28i2 

 in this line is within the area of the East Wind Drift. The line in 0° is rather longer than that in 90° E, 

 and it has some rather peculiar physical characteristics. The subantarctic zone extends from a fairly sharp 

 subtropical convergence in 39" S to a rather diffuse antarctic convergence in about 50° S. The antarctic 

 zone was accessible thence to nearly 70° S. This zone is, however, complicated by the presence of the 

 Weddell Drift, a broad tongue of cold water originating in the Weddell Sea and deflected by the coast 

 of Grahamland in a north and easterly direction. Deacon (1936) shows its northern limit to be in 

 about 52° S in 0° and its southern boundary is probably in about 61° S, though it is difficult to define 

 accurately, and both limits indeed are no doubt to be taken as tentative and subject to variation. 

 Stations 2010 and probably 2012 lie within the influence of the East Wind Drift, and it is possible 

 that there may be swirls or eddies between the East Wind Drift and the Weddell Current (Marr, 

 personal communication). The line in 80° W is the shortest of the three, and the surface isotherms 

 show that considerable north/south compression of the water masses takes place in this area. The 

 width of the subantarctic zone is not known, since the position of the subtropical convergence has 



